BHG Infra
Project tabs

Knowledge tab

A project's documents. Add meeting notes, transcripts, emails, SOWs and BRDs here, read each in three layers, and Sync reads them into proposed changes.

The Knowledge tab holds everything written about the project: meeting notes, transcripts, emails, the SOW, the BRD and other documents. It is the project's source material. Sync reads from here, the project agent searches here, and every proposed change points back to a document here.

The layout

  • On the left, the document list. Search with Search this project's documents…, filter by type, and open a document. Documents are grouped into sections.
  • On the right, the reader. You can open several documents as tabs.
  • The footer shows how many documents there are, and Add if you can write.

On a project, questions about the documents go to the project agent. Use Ask the Project agent on a document to ask the agent about it.

The three layers of a document

A document can have up to three layers. The switch above the document shows only the layers it has:

LayerWhat it is
OriginalThe document as it was added.
TranscriptThe meeting transcript, when one was stored with it. Granola and Teams transcripts appear here when the connector or the uploading agent sends one.
SummaryThe facts read from the document, grouped under headings such as Decisions and answers, Open questions, Actions, Risks, issues and dependencies and Requirements. Each fact shows its quote.

The summary comes from one of two places:

  • What Infra read from this document: the facts Sync found.
  • Summary by an agent: the digest an agent sent when it staged the document. See Staging documents with a digest.

A quote Infra could not find word for word in the document is marked quote not found in the document. A fact with no quote reads "No quote: as stated".

Add a document

Click Add in the footer of the document list.

  1. Pick the Type: Meeting note, Email, SOW, Change order, BRD, Table or Other. The type tells Sync how to read it.
  2. Either Upload Markdown (a .md file, 500 KB max), or paste the text: an Artifact title, the kind (Note, Transcript, Document or Email), the source (Pasted, Granola, Teams or Upload), the text, and an optional Source URL.
  3. Click Add + index.

You can also drop files onto the Knowledge tab. Dropped files can be .md, .txt, .csv, .json, .eml, .docx, .pdf or .xlsx. Text files may be up to 500 KB; .docx, .pdf, .xlsx and .csv files up to 5 MB. Infra converts them to text. Old .xls files are refused: save them as .xlsx first.

A new document is indexed so it can be searched. If indexing fails, the document is kept and you can Retry indexing. Archive removes a document from search and from the agent.

Bring in meetings, emails and files with connectors

Two connectors let the project agent copy sources onto a project for you. Each works as you, with your own account.

  • Granola: meeting notes and transcripts from your Granola account.
  • Microsoft 365: Outlook email, your calendar, Teams meeting transcripts, and SharePoint or OneDrive files you can open.

Connect them once on Knowledge Base › Connections › Connectors to Infra: click Connect and sign in. Then ask the project agent, for example "Stage yesterday's Acme status call from Granola" or "Find the signed SOW in SharePoint and add it". The agent finds the source, shows what it found and copies it onto the project as a document. It does not Sync it: you review it with Sync.

A file from SharePoint or OneDrive can be .docx, .pdf, .xlsx, .csv, .md or .txt.

Read a document again

Sync reads each document once, and again when it changes. To read one document again on purpose (for example a SOW or BRD after the reader improved), open it and click Read again. Its proposed changes go to Events like any Sync.

On this page